2.073 - fix behavior and tests for openssl 3.0.1 - fix #110 - prevent internal error warning in some cases 2.072 - add PEM_certs2file and PEM_file2certs in IO::Socket::SSL::Utils based on idea by rovo89 in #101 - certs/*.p12 used for testing should now work with OpenSSL 3.0 too #108 - update public suffix database 2.071 2021/05/23 - fix t/nonblock.t race on some systems. Fixes issue #102, maybe #98 too. 2.070 2021/02/26 - changed bugtracker in Makefile.PL to github, away from obsolete rt.cpan.org 2.069 2021/01/22 - IO::Socket::Utils CERT_asHash and CERT_create now support subject and issuer with multiple same parts (like multiple OU). In this case an array ref instead of a scalar is used as hash value. https://github.com/noxxi/p5-io-socket-ssl/issues/95 2.068 2020/03/31 - treat OpenSSL 1.1.1e as broken and refuse to build with it in order to prevent follow-up problems in tests and user code https://github.com/noxxi/p5-io-socket-ssl/issues/93 https://github.com/openssl/openssl/issues/11388 https://github.com/openssl/openssl/issues/11378 - update PublicSuffix with latest data from publicsuffix.org 2.067 2020/02/14 - fix memory leak on incomplete handshake https://github.com/noxxi/p5-io-socket-ssl/issues/92 Thanks to olegwtf - add support for SSL_MODE_RELEASE_BUFFERS via SSL_mode_release_buffers This can decrease memory usage at the costs of more allocations https://rt.cpan.org/Ticket/Display.html?id=129463 - more detailed error messages when loading of certificate file failed https://github.com/noxxi/p5-io-socket-ssl/issues/89 - fix for ip_in_cn == 6 in verify_hostname scheme https://rt.cpan.org/Ticket/Display.html?id=131384 - deal with new MODE_AUTO_RETRY default in OpenSSL 1.1.1 - fix warning when no ecdh support is available - documentation update regarding use of select and TLS 1.3 - various fixes in documentation https://github.com/noxxi/p5-io-socket-ssl/issues/91 https://github.com/noxxi/p5-io-socket-ssl/issues/90 https://github.com/noxxi/p5-io-socket-ssl/issues/87 https://github.com/noxxi/p5-io-socket-ssl/issues/81 - stability fix t/core.t 2.066 2019/03/06 - fix test t/verify_partial_chain.t by using the newly exposed function can_partial_chain instead of guessing (wrongly) if the functionality is available 2.065 2019/03/05 - make sure that Net::SSLeay::CTX_get0_param is defined before using X509_V_FLAG_PARTIAL_CHAIN. Net::SSLeay 1.85 defined only the second with LibreSSL 2.7.4 but not the first https://rt.cpan.org/Ticket/Display.html?id=128716 - prefer AES for server side cipher default since it is usually hardware-accelerated 2.064 2019/03/04 - make algorithm for fingerprint optional, i.e. detect based on length of fingerprint - https://rt.cpan.org/Ticket/Display.html?id=127773 - fix t/sessions.t and improve stability of t/verify_hostname.t on windows - use CTX_set_ecdh_auto when needed (OpenSSL 1.0.2) if explicit curves are set - update fingerprints for live tests 2.063 2019/03/01 - support for both RSA and ECDSA certificate on same domain - update PublicSuffix - Refuse to build if Net::SSLeay is compiled with one version of OpenSSL but then linked against another API-incompatible version (ie. more than just the patchlevel differs). 2.062 2019/02/24 - Enable X509_V_FLAG_PARTIAL_CHAIN if supported by Net::SSLeay (1.83+) and OpenSSL (1.1.0+). This makes leaf certificates or intermediate certificates in the trust store be usable as full trust anchors too. 2.061 2019/02/23 - Support for TLS 1.3 session reuse. Needs Net::SSLeay 1.86+. Note that the previous (and undocumented) API for the session cache has been changed. - Support for multiple curves, automatic setting of curves and setting of supported curves in client. Needs Net::SSLeay 1.86+. - Enable Post-Handshake-Authentication (TLSv1.3 feature) client-side when client certificates are provided. Thanks to jorton[AT]redhat[DOT]com. Needs Net::SSLeay 1.86+. 2.060 2018/09/16 - support for TLS 1.3 with OpenSSL 1.1.1 (needs Net::SSLeay 1.86+) Thanks to ppisar[AT]redhat.com for major help see also https://rt.cpan.org/Ticket/Display.html?id=126899 TLS 1.3 support is not complete yet for session reuse 2.059 2018/08/15 - fix memleak when CRL are used. Thanks to Franz Skale for report and patch https://rt.cpan.org/Ticket/Display.html?id=125867 - fix memleak when using stop_SSL and threads, reported by Paul Evans https://rt.cpan.org/Ticket/Display.html?id=125867#txn-1797132 2.058 2018/07/19 - fix t/session_ticket.t: it failed with OpenSSL 1.1.* since this version expects the extKeyUsage of clientAuth in the client cert also to be allowed by the CA if CA uses extKeyUsage 2.057 2018/07/18 - fix memory leak which occurred with explicit stop_SSL in connection with non-blocking sockets or timeout - https://rt.cpan.org/Ticket/Display.html?id=125867 Thanks to Paul Evans for reporting - fix redefine warnings in case Socket6 is installed but neither IO::Socket::IP nor IO::Socket::INET6 - https://rt.cpan.org/Ticket/Display.html?id=124963 - IO::Socket::SSL::Intercept - optional 'serial' argument can be starting number or callback to create serial number based on the original certificate - new function get_session_reused to check if a session got reused - IO::Socket::SSL::Utils::CERT_asHash: fingerprint_xxx now set to the correct value 2.056 2018/02/19 - Intercept - fix creation of serial number: base it on binary digest instead of treating hex fingerprint as binary. Allow use of own serial numbers again. - t/io-socket-ip.t - skip test if no IPv6 support on system RT#124464 - update PublicSuffix 2.055 2018/02/15 - use SNI also if hostname was given all-uppercase - Utils::CERT_create - don't add authority key for issuer since Chrome does not like this - Intercept: - change behavior of code based cache to better support synchronizing within multiprocess/threaded setups - don't use counter for serial number but somehow base it on original certificate in order to avoid conflicts with reuse of serial numbers after restart - RT#124431 - better support platforms w/o IPv6 - RT#124306 - spelling fixes in documentation 2.054 2018/01/22 - added missing test certificates to MANIFEST 2.053 2018/01/21 - small behavior fixes - if SSL_fingerprint is used and matches don't check for OCSP - Utils::CERT_create - small fixes to properly specific purpose, ability to use predefined complex purpose but disable some features - update PublicSuffix - updates for documentation, especially regarding pitfalls with forking or using non-blocking sockets. Spelling fixes. - test fixes and improvements - stability improvements for live tests - regenerate certificate in certs/ and make sure they are limited to the correct purpose. Checkin program used to generate certificates. - adjust tests since certificates have changed and some tests used certificates intended for client authentication as server certificates, which now no longer works 2.052 2017/10/22 - disable NPN support if LibreSSL>=2.6.1 is detected since they've replaced the functions with dummies instead of removing NPN completly or setting OPENSSL_NO_NEXTPROTONEG - t/01loadmodule.t shows more output helpful in debugging problems - update fingerprints for extenal tests - update documentation to make behavior of syswrite more clear 2.051 2017/09/05 - syswrite: if SSL_write sets SSL_ERROR_SYSCALL but no $! (as seen with OpenSSL 1.1.0 on Windows) set $! to EPIPE to propagate a useful error up https://github.com/noxxi/p5-io-socket-ssl/issues/62 2.050 2017/08/18 - removed unnecessary settings of SSL_version and SSL_cipher_list from tests - protocol_version.t can now deal when TLS 1.0 and/or TLS 1.1 are not supported as is the case with openssl versions in latest Debian (buster) 2.049 2017/06/12 - fixed problem caused by typo in the context of session cache https://github.com/noxxi/p5-io-socket-ssl/issues/60 - update PublicSuffix information from publicsuffix.org 2.048 2017/04/16 - fixed small memory leaks during destruction of socket and context, RT#120643 2.047 2017/02/16 - better fix for problem which 2.046 tried to fix but broke LWP this way 2.046 2017/02/15 - cleanup everything in DESTROY and make sure to start with a fresh %{*self} in configure_SSL because it can happen that a GLOB gets used again without calling DESTROY (https://github.com/noxxi/p5-io-socket-ssl/issues/56) 2.045 2017/02/13 - fixed memory leak caused by not destroying CREATED_IN_THIS_THREAD for SSL objects -> github pull#55 - optimization: don't track SSL objects and CTX in *CREATED_IN_THIS_THREAD if perl is compiled w/o thread support - small fix in t/protocol_version.t to use older versions of Net::SSLeay with openssl build w/o SSLv3 support - when setting SSL_keepSocketOnError to true the socket will not be closed on fatal error. This is a modified version of https://github.com/noxxi/p5-io-socket-ssl/pull/53/ 2.044 2017/01/26 - protect various 'eval'-based capability detections at startup with a localized __DIE__ handler. This way dynamically requiring IO::Socket::SSL as done by various third party software should cause less problems even if there is a global __DIE__ handler which does not properly deal with 'eval'. 2.043 2017/01/06 - make t/session_ticket.t work with OpenSSL 1.1.0. With this version the session does not get reused any longer if it was not properly closed which is now done using an explicit close by the client which causes a proper SSL_shutdown 2.042 2017/01/05 - enable session ticket callback with Net::SSLeay>=1.80 2.041 2017/01/04 - leave session ticket callback off for now until the needed patch is included in Net::SSLeay. See https://rt.cpan.org/Ticket/Display.html?id=116118#txn-1696146 2.040 2016/12/17 - fix detection of default CA path for OpenSSL 1.1.x - Utils::CERT_asHash now includes the signature algorithm used - Utils::CERT_asHash can now deal with large serial numbers 2.039 2016/11/20 - OpenSSL 1.1.0c changed the behavior of SSL_read so that it now returns -1 on EOF without proper SSL shutdown. Since it looks like that this behavior will be kept at least for 1.1.1+ adapt to the changed API by treating errno=NOERR on SSL_ERROR_SYSCALL as EOF. 2.038 2016/09/17 - restrict session ticket callback to Net::SSLeay 1.79+ since version before contains bug. Add test for session reuse - extend SSL fingerprint to pubkey digest, i.e. 'sha1$pub$xxxxxx....' - fix t/external/ocsp.t to use different server (under my control) to check OCSP stapling 2.037 2016/08/22 - fix session cache del_session: it freed the session but did not properly remove it from the cache. Further reuse causes crash. 2.036 2016/08/11 - disable OCSP support when Net::SSLeay 1.75..1.77 is used, see RT#116795 2.035 2016/08/11 - fixes for issues introduced in 2.034 - return with error in configure_SSL if context creation failed. This might otherwise result in a segmentation fault later. - apply builtin defaults before any (user configurable) global settings (i.e. done with set_defaults, set_default_context...) so that builtins don't replace user settings Thanks to joel[DOT]a[DOT]berger[AT]gmail[DOT]com for reporting 2.034 2016/08/08 - move handling of global SSL arguments into creation of context, so that these get also applied when creating a context only. 2.033 2016/07/15 - support for session ticket reuse over multiple contexts and processes (if supported by Net::SSLeay) - small optimizations, like saving various Net::SSLeay constants into variables and access variables instead of calling the constant sub all the time - make t/dhe.t work with openssl 1.1.0 2.032 2016/07/12 - Set session id context only on the server side. Even if the documentation for SSL_CTX_set_session_id_context makes clear that this function is server side only it actually affects hndling of session reuse on the client side too and can result in error "SSL3_GET_SERVER_HELLO:attempt to reuse session in different context" at the client. 2.031 2016/07/08 - fix for bug in session handling introduced in 2.031, RT#115975 Thanks to paul[AT]city-fan[DOT]org for reporting 2.030 2016/07/08 - Utils::CERT_create - don't add given extensions again if they were already added. Firefox croaks with sec_error_extension_value_invalid if (specific?) extensions are given twice. - assume that Net::SSLeay::P_PKCS12_load_file will return the CA certificates with the reverse order as in the PKCS12 file, because that's what it does. - support for creating ECC keys in Utils once supported by Net::SSLeay - remove internal sub session_cache and access cache directly (faster) 2.029 2016/06/26 - fix del_session method in case a single item was in the cache - use SSL_session_key as the real key for the cache and not some derivative of it, so that it works to remove the entry using the same key 2.028 2016/06/26 - add del_session method to session cache 2.027 2016/04/20 - only added Changes for 2.026 2.026 2016/04/20 - update default server and client ciphers based on recommendation of Mozilla and what the current browsers use. Notably this finally disables RC4 for the client (was disabled for server long ago) and adds CHACHA20. 2.025 2016/04/04 - Resolved memleak if SSL_crl_file was used: RT#113257, RT#113530 Thanks to avi[DOT]maslati[AT]forescout[DOT]com and mark[DOT]kurman[AT]gmail[DOT]com for reporting the problem 2.024 2016/02/06 - Work around issue where the connect fails on systems having only a loopback interface and where IO::Socket::IP is used as super class (default when available). Since IO::Socket::IP sets AI_ADDRCONFIG by default connect to localhost would fail on this systems. This happened at least for the tests, see https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813796 Workaround is to explicitly set GetAddrInfoFlags to 0 if no GetAddrInfoFlags is set but the Family/Domain is given. In this case AI_ADDRCONFIG would not be useful anyway but would cause at most harm. 2.023 2016/01/30 - OpenSSL 1.0.2f changed the behavior of SSL shutdown in case the TLS connection was not fully established (commit: f73c737c7ac908c5d6407c419769123392a3b0a9). This somehow resulted in Net::SSLeay::shutdown returning 0 (i.e. keep trying) which caused an endless loop. It will now ignore this result in case the TLS connection was not yet established and consider the TLS connection closed instead. 2.022 2015/12/10 - fix stringification of IPv6 inside subjectAltNames in Utils::CERT_asHash. Thanks to Mark.Martinec[AT]ijs[DOT]si for reporting in #110253 2.021 2015/12/02 - Fixes for documentation and typos thanks to DavsX and jwilk. - Update PublicSuffix with latest version from publicsuffix.org 2.020 2015/09/20 - support multiple directories in SSL_ca_path as proposed in RT#106711 by dr1027[AT]evocat[DOT]ne. Directories can be given as array or as string with a path separator, see documentation. - typos fixed thanks to jwilk https://github.com/noxxi/p5-io-socket-ssl/pull/34 2.019 2015/09/01 - work around different behavior of getnameinfo from Socket and Socket6 by using a different wrapper depending on which module I use for IPv6. Thanks to bluhm for reporting. 2.018 2015/08/27 - RT#106687 - startssl.t failed on darwin with old openssl since server requested client certificate but offered also anon ciphers 2.017 2015/08/24 - checks for readability of files/dirs for certificates and CA no longer use -r because this is not safe when ACLs are used. Thanks to BBYRD, RT#106295 - new method sock_certificate similar to peer_certificate based on idea of Paul Evans, RT#105733 - get_fingerprint can now take optional certificate as argument and compute the fingerprint of it. Useful in connection with sock_certificate. - check for both EWOULDBLOCK and EAGAIN since these codes are different on some platforms. Thanks to Andy Grundman, RT#106573 - enforce default verification scheme if none was specified, i.e. no longer just warn but accept. If really no verification is wanted a scheme of 'none' must be explicitly specified. - support different cipher suites per SNI hosts 2.016 2015/06/02 - add flag X509_V_FLAG_TRUSTED_FIRST by default if available in OpenSSL (since 1.02) and available with Net::SSLeay. RT#104759 (thanks GAAS) - work around hanging prompt() with older perl in Makefile.PL RT#104731 - make t/memleak_bad_handshake.t work on cygwin and other systems having /proc/pid/statm, see RT#104659 - add better debugging based on patch from H.Merijn Brand 2.015 2015/05/13 - work around problem with IO::Socket::INET6 on windows, by explicitly using Domain AF_INET in the tests. Fixes RT#104226 reported by CHORNY 2.014 2015/05/05 - Utils::CERT_create - work around problems with authorityInfoAccess, where OpenSSL i2v does not create the same string as v2i expects - Intercept - don't clone some specific extensions which make only sense with the original certificate 2.013 2015/05/01 - assign severities to internal error handling and make sure that follow-up errors like "configuration failed" or "certificate verify error" don't replace more specific "hostname verification failed" when reporting in sub errstr/$SSL_ERROR. see also RT#103423 - enhanced documentation thanks to Chase Whitener https://github.com/noxxi/p5-io-socket-ssl/pull/26 2.012 2015/02/02 - fix t/ocsp.t in case no HTTP::Tiny is installed 2.011 2015/02/01 - fix t/ocsp.t - don't count on revoked.grc.com using OCSP stapling #101855 - added option 'purpose' to Utils::CERT_create to get better control of the certificates purpose. Default is 'server,client' for non-CA (contrary to only 'server' before) - removed RC4 from default cipher suites on the server site https://github.com/noxxi/p5-io-socket-ssl/issues/22 - refactoring of some tests using Test::More thanks to Sweet-kid and the 2015 Pull Request Challenge 2.010 2015/01/14 - new options SSL_client_ca_file and SSL_client_ca to let the server send the list of acceptable CAs for the client certificate. - t/protocol_version.t - fix in case SSLv3 is not supported in Net::SSLeay. RT#101485, thanks to TEAM. 2.009 2015/01/12 - remove util/analyze.pl. This tool is now together with other SSL tools in https://github.com/noxxi/p5-ssl-tools - added ALPN support (needs OpenSSL1.02, Net::SSLeay 1.56+) thanks to TEAM, RT#101452 2.008 2014/12/16 - work around recent OCSP verification errors for revoked.grc.com (badly signed OCSP response, Firefox also complains about it) in test t/external/ocsp.t. - util/analyze.pl - report more details about preferred cipher for specific TLS versions 2.007 2014/11/26 - make getline/readline fall back to super class if class is not sslified yet, i.e. behave the same as sysread, syswrite etc. This fixes RT#100529 2.006 2014/11/22 - Make (hopefully) non-blocking work on windows by using EWOULDBLOCK instead of EAGAIN. While this is the same on UNIX it is different on Windows and socket operations return there (WSA)EWOULDBLOCK and not EAGAIN. Enable non-blocking tests on Windows too. - make PublicSuffix::_default_data thread safe - update PublicSuffix with latest list from publicsuffix.org 2.005 2014/11/15 - next try to fix t/protocol_version.t for OpenSSL w/o SSLv3 support 2.004 2014/11/15 - only test fix: fix t/protocol_version.t to deal with OpenSSL installations which are compiled without SSLv3 support. 2.003 2014/11/14 - make SSLv3 available even if the SSL library disables it by default in SSL_CTX_new (like done in LibreSSL). Default will stay to disable SSLv3, so this will be only done when setting SSL_version explicitly. - fix possible segmentation fault when trying to use an invalid certificate, reported by Nick Andrew. - Use only the ICANN part of the default public suffix list and not the private domains. This makes existing exceptions for s3.amazonaws.com and googleapis.com obsolete. Thanks to Gervase Markham from mozilla.org. 2.002 2014/10/21 - fix check for (invalid) IPv4 when validating hostname against certificate. Do not use inet_aton any longer because it can cause DNS lookups for malformed IP. RT#99448, thanks to justincase[AT]yopmail[DOT]com. - Update PublicSuffix with latest version from publicsuffix.org - lots of new top level domains. - Add exception to PublicSuffix for s3.amazonaws.com - RT#99702, thanks to cpan[AT]cpanel[DOT]net. 2.001 2014/10/21 - Add SSL_OP_SINGLE_(DH|ECDH)_USE to default options to increase PFS security. Thanks to Heikki Vatiainen for suggesting. - Update external tests with currently expected fingerprints of hosts. - Some fixes to make it still work on 5.8.1. 2.000 2014/10/15 - consider SSL3.0 as broken because of POODLE and disable it by default. - Skip live tests without asking if environment NO_NETWORK_TESTING is set. Thanks to ntyni[AT]debian[DOT]org for suggestion. - skip tests which require fork on non-default windows setups without proper fork. Thanks to SHAY for https://github.com/noxxi/p5-io-socket-ssl/pull/18 1.999 2014/10/09 - make sure we don't use version 0.30 of IO::Socket::IP - make sure that PeerHost is checked on all places where PeerAddr is checked, because these are synonyms and IO::Socket::IP prefers PeerHost while others prefer PeerAddr. Also accept PeerService additionally to PeerPort. See https://github.com/noxxi/p5-io-socket-ssl/issues/16 for details. - add ability to use client certificates and to overwrite hostname with util/analyze-ssl.pl. 1.998 2014/09/07 - make client authentication work at the server side when SNI is in by use having CA path and other settings in all SSL contexts instead of only the main one. Based on code from lundstrom[DOT]jerry[AT]gmail[DOT]com, https://github.com/noxxi/p5-io-socket-ssl/pull/15 1.997 2014/07/12 - thanks to return code 1 from Net::SSLeay::library_init if the library needed initialization and 0 if not we can now clearly distinguish if initialization was needed and do not need any work-arounds for perlcc by the user. 1.996 2014/07/12 - move initialization of OpenSSL-internals out of INIT again because this breaks if module is used with require. Since there is no right place to work in all circumstances just document the work-arounds needed for perlcc. RT#97166 1.995 2014/07/11 - RT#95452 - move initialization and creation of OpenSSL-internals into INIT section, so they get executed after compilation and perlcc is happy. - refresh option for peer_certificate, so that it checks if the certificate changed in the mean time (on renegotiation) - fix fingerprint checking - now applies only to topmost certificate - IO::Socket::SSL::Utils - accept extensions within CERT_create - documentations fixes thanks to frioux - fix documentation bug RT#96765, thanks to Salvatore Bonaccorso. 1.994 2014/06/22 - IO::Socket::SSL can now be used as dual-use socket, e.g. start plain, upgrade to SSL and downgrade again all with the same object. See documentation of SSL_startHandshake and chapter Advanced Usage. - try to apply SSL_ca* even if verify_mode is 0, but don't complain if this fails. This is needed if one wants to explicitly verify OCSP lookups even if verification is otherwise off, because otherwise the signature check would fail. This is mostly useful for testing. - reorder documentation of attributes for new, so that the more important ones are at the top. 1.993 2014/06/13 - major rewrite of documentation, now in separate file - rework error handling to distinguish between SSL errors and internal errors (like missing capabilities). - fix handling of default_ca if given during the run of the program (Debian#750646) - util/analyze-ssl.pl - fix hostname check if SNI does not work 1.992 2014/06/01 - set $! to undef before doing IO (accept, read..). On Windows a connection reset could cause SSL read error without setting $!, so make sure we don't keep the old value and maybe thus run into endless loop. 1.991 2014/05/27 - new option SSL_OCSP_TRY_STAPLE to enforce staple request even if VERIFY_NONE - work around for RT#96013 in peer_certificates 1.990 2014/05/27 - added option SSL_ocsp_staple_callback to get the stapled OCSP response and verify it somewhere else - try to fix warnings on Windows again (#95967) - work around temporary OCSP error in t/external/ocsp.t 1.989 2014/05/24 - fix #95881 (warnings on windows), thanks to TMHALL 1.988 2014/05/17 - add transparent support for DER and PKCS#12 files to specify cert and key, e.g. it will autodetect the format - if SSL_cert_file is PEM and no SSL_key_file is given it will check if the key is in SSL_cert_file too 1.987 2014/05/17 - fix t/verify_hostname_standalone.t on systems without usable IDNA or IPv6 #95719, thanks srchulo - enable IPv6 support only if we have a usable inet_pton - remove stale entries from MANIFEST (thanks seen[AT]myfairpoint[DOT]net) 1.986 2014/05/16 - allow IPv4 in common name, because browsers allow this too. But only for scheme www/http, not for rfc2818 (because RC2818 does not allow this). In default scheme IPv6 and IPv4 are allowed in CN. Thanks to heiko[DOT]hund[AT]sophos[DOT]com for reporting the problem. - Fix handling of public suffix. Add exemption for *.googleapis.com wildcard, which should be better not allowed according to public suffix list but actually is used. - Add hostname verification test based on older test of chromium. But change some of the test expectations because we don't want to support IP as SAN DNS and because we enforce a public suffix list (and thus *.co.uk should not be allowed) 1.985 2014/05/15 - make OCSP callback return 1 even if it was called on the server side because of bad setup of the socket. Otherwise we get an endless calling of the OCSP callback. - consider an OCSP response which is not yet or no longer valid a soft error instead of a hard error - fix skip in t/external/ocsp.t in case fingerprint does not match - RT#95633 call EVP_PKEY_free not EVP_KEY_free in IO::Socket::SSL::Utils::KEY_free. Thanks to paul[AT]city-fan[DOT]org - util/analyze.pl - with --show-chain check if chain with SNI is different from chain w/o SNI. 1.984 2014/05/10 - added OCSP support: - needs Net::SSLeay >=1.59 - for usage see documentation of IO::Socket::SSL (examples and anything with OCSP in the name) - new tool util/analyze-ssl.pl which is intended to help in debugging of SSL problems and to get information about capabilities of server. Works also as en example of how to use various features (like OCSP, SNI..) - fix peer_certificates (returns leaf certificate only once on client side) - added timeout for stop_SSL (either with Timeout or with the default timeout for IO::Socket) - fix IO::Socket::SSL::Utils mapping between ASN1_TIME and time_t when local time is not GMT. Use Net::SSLeay::ASN1_TIME_timet if available. - fix t/external/usable_ca.t for system with junk in CA files 1.983 2014/05/03 - fix public suffix handling: ajax.googleapis.com should be ok even if googleapis.com is in public suffix list (e.g. check one level less) #95317, thanks to purification[AT]ukr[DOT]net - usable_ca.t - update fingerprints after heartbleed attack - usable_ca.t - make sure we have usable CA for tested hosts in CA store 1.982 2014/04/24 - fix for using subroutine as argument to set_args_filter_hack 1.981 2014/04/08 - #95432 fix ecdhe Test for openssl1.0.1d, thanks to paul[AT]city-fan[DOT]org - fix detection of openssl1.0.1d (detected 1.0.1e instead) - new function can_ecdh in IO::Socket::SSL 1.980 2014/04/08 - fixed incorrect calculation of certificate fingerprint in get_fingerprint* and comparison in SSL_fingerprint. Thanks to david[DT]palmer[AT]gradwell[DOT]com for reporting. - disable elliptic curve support for openssl 1.0.1d on 64bit because of openssl rt#2975 1.979 2014/04/06 - hostname checking: - configuration of 'leftmost' is renamed to 'full_label', but the old version is kept for compatibility reasons. - documentation of predefined schemes fixed to match reality 1.978 2014/04/04 - RT#94424 again, fix test on older openssl version with no SNI support 1.977 2014/04/04 - fix publicsuffix for IDNA, more tests with various IDNA libs RT#94424. Thanks to paul[AT]city-fan[DOT]org - reuse result of IDN lib detection from PublicSuffix.pm in SSL.pm - add more checks to external/usable_ca.t. Now it is enough that at least one of the hosts verifies against the builtin CA store - add openssl and Net::SSleay version to diagnostics in load test 1.976 2014/04/03 - added public prefix checking to verification of wildcard certificates, e.g. accept *.foo.com but not *.co.uk. See documentation of SSL_verifycn_publicsuffix and IO::Socket::SSL::PublicSuffix Thanks to noloader for pointing out the problem. 1.975 2014/04/02 - BEHAVIOR CHANGE: work around TEA misfeature on OS X builtin openssl, e.g. guarantee that only the explicitly given CA or the openssl default CA will be used. This means that certificates inside the OS X keyring will no longer be used, because there is no way to control the use by openssl (e.g. certificate pinning etc) - make external tests run by default to make sure default CA works on all platforms, it skips automatically on network problems like timeouts or ssl interception, can also use http(s)_proxy environment variables 1.974 2014/04/02 - new function peer_certificates to get the whole certificate chain, needs Net::SSLeay>=1.58 - extended IO::Socket::Utils::CERT_asHash to provide way more information, like issuer information, cert and pubkey digests, all extensions, CRL distributions points and OCSP uri 1.973 2014/03/25 - with SSL_ca certificate handles can now be used additionally to SSL_ca_file and SSL_ca_path - do not complain longer if SSL_ca_file and SSL_ca_path are both given, instead add both as options to the CA store - Shortcut 'issuer' to give both issuer_cert and issuer_key in CERT_create. 1.972 2014/03/23 - make sure t/external/usable_ca.t works also with older openssl without support for SNI. RT#94117. Thanks to paul[AT]city-fan[DOT]org 1.971 2014/03/22 - try to use SSL_hostname for hostname verification if no SSL_verifycn_name is given. This way hostname for SNI and verification can be specified in one step. - new test program example/simulate_proxy.pl 1.970 2014/03/19 - fix rt#93987 by making sure sub default_ca does use a local $_ and not a version of an outer scope which might be read-only. Thanks to gshank 1.969 2014/03/13 - fix set_defaults to match documentation regarding short names - new function set_args_filter_hack to make it possible to override bad SSL settings from other code at the last moment. - determine default_ca on module load (and not on first use in each thread) - don't try default hostname verification if verify_mode 0 - fix hostname verification when reusing context 1.968 2014/03/13 - BEHAVIOR CHANGE: removed implicit defaults of certs/server-{cert,key}.pem for SSL_{cert,key}_file and ca/,certs/my-ca.pem for SSL_ca_file. These defaults were depreceated since 1.951 (2013/7/3). - Usable CA verification path on Windows etc: Do not use Net::SSLeay::CTX_set_default_verify_paths any longer to set system/build dependent default verification path, because there was no way to retrieve these default values and check if they contained usable CA. Instead re-implement the same algorithm and export the results with public function default_ca() and make it possible to overwrite it. Also check for usable verification path during build. If no usable path are detected require Mozilla::CA at build and try to use it at runtime. 1.967 2014/02/06 - verify the hostname inside a certificate by default with a superset of common verification schemes instead of not verifying identity at all. For now it will only complain if name verification failed, in the future it will fail certificate verification, forcing you to set the expected SSL_verifycn_name if you want to accept the certificate. - new option SSL_fingerprint and new methods get_fingerprint and get_fingerprint_bin. Together they can be used to selectively accept specific certificates which would otherwise fail verification, like self-signed, outdated or from unknown CAs. This makes another reason to disable verification obsolete. - Utils: - default RSA key length 2048 - digest algorithm to sign certificate in CERT_create can be given, defaults to SHA-256 - CERT_create can now issue non-CA selfsigned certificate - CERT_create add some more useful constraints to certificate - spelling fixes, thanks to ville[dot]skytta[at]iki[dot]fi 1.966 2014/01/21 - fixed bug introduced in 1.964 - disabling TLSv1_2 worked no longer with specifying !TLSv12, only !TLSv1_2 worked - fixed leak of session objects in SessionCache, if another session replaced an existing session (introduced in 1.965) 1.965 2014/01/16 - new key SSL_session_key to influence how sessions are inserted and looked up in the clients session cache. This makes it possible to share sessions over different ip:host (like required with some FTPS servers) - t/core.t - handle case, were default loopback source is not 127.0.0.1, like in FreeBSD jails 1.964 2014/01/15 - Disabling TLSv1_1 did not work, because the constant was wrong. Now it gets the constants from calling Net::SSLeay::SSL_OP_NO_TLSv1_1 etc - The new syntax for the protocols is TLSv1_1 instead of TLSv11. This matches the syntax from OpenSSL. The old syntax continues to work in SSL_version. - New functions get_sslversion and get_sslversion_int which get the SSL version of the establish session as string or int. - disable t/io-socket-inet6.t if Acme::Override::INET is installed 1.963 2014/01/13 - fix behavior of stop_SSL: for blocking sockets it now enough to call it once, for non-blocking it should be called again as long as EAGAIN and SSL_ERROR is set to SSL_WANT_(READ|WRITE). - don't call blocking if start_SSL failed and downgraded socket has no blocking method, thanks to tokuhirom - documentation enhancements: - special section for differences to IO::Socket - describe problem with blocking accept on non-blocking socket - describe arguments to new_from_fd and make clear, that for upgrading an existing IO::Socket start_SSL should be used directly 1.962 2013/11/27 - work around problems with older F5 BIG-IP by offering fewer ciphers on the client side by default, so that the client hello stays below 255 byte 1.961 2013/11/26 - IO::Socket::SSL::Utils::CERT_create can now create CA-certificates which are not self-signed (by giving issuer_*) 1.960 2013/11/12 only documentation enhancements: - clarify with text and example code, that within event loops not only select/poll should be used, but also pending has to be called. - better introduction into SSL, at least mention anonymous authentication as something you don't want and should take care with the right cipher - make it more clear, that user better does not change the cipher list, unless he really know what he is doing 1.959 2013/11/12 - bugfix test core.t windows only 1.958 2013/11/11 - cleanup: remove workaround for old IO::Socket::INET6 but instead require at least version 2.55 which is now 5 years old - fix t/session.t #RT90240, thanks to paul[AT]city-fan[DOT]org 1.957 2013/11/11 - fixed t/core.t: test uses cipher_list of HIGH, which includes anonymous authorization. With the DH param given by default since 1.956 old versions of openssl (like 0.9.8k) used cipher ADH-AES256-SHA (e.g. anonymous authorization) instead of AES256-SHA and thus the check for the peer certificate failed (because ADH does not exchanges certificates). Fixed by explicitly specifying HIGH:!aNULL as cipher RT#90221, thanks to paul[AT]city-fan[DOT]org - cleaned up tests: - remove ssl_settings.req and 02settings.t, because all tests now create a simple socket at 127.0.0.1 and thus global settings are no longer needed. - some tests did not have use strict(!), fixed it. - removed special handling for older Net::SSLeay versions, which are less than our minimum requirement - some syntax enhancements, removed some SSL_version and SSL_cipher_list options where they were not really needed 1.956 2013/11/10 lots of behavior changes for more secure defaults: - BEHAVIOR CHANGE: make default cipher list more secure, especially - no longer support MD5 by default (broken) - no longer support anonymous authentication by default (vulnerable to man in the middle attacks) - prefer ECDHE/DHE ciphers and add necessary ECDH curve and DH keys, so that it uses by default forward secrecy, if underlying Net::SSLeay/openssl supports it - move RC4 at the end, e.g. 3DES is preferred (BEAST attack should hopefully been fixed and now RC4 is considered less safe than 3DES) - default SSL_honor_cipher_order to 1, e.g. when used as server it tries to get the best cipher even if client prefers other ciphers PLEASE NOTE that this might break connections with older, less secure implementations. In this case revert to 'ALL:!LOW:!EXP:!aNULL' or so. - BEHAVIOR CHANGE: SSL_cipher_list now gets set on context not SSL object and thus gets reused if context gets reused. PLEASE NOTE that using SSL_cipher_list together with SSL_reuse_ctx has no longer effect on the ciphers of the context. - rework hostname verification schemes - add rfc names as scheme (e.g. 'rfc2818',...) - add SIP, SNMP, syslog, netconf, GIST - BEHAVIOR CHANGE: fix SMTP - now accept wildcards in CN and subjectAltName - BEHAVIOR CHANGE: fix IMAP, POP3, ACAP, NNTP - now accept wildcards in CN - BEHAVIOR CHANGE: anywhere wildcards like www* now match only 'www1', 'www2'.. but not 'www' - anywhere wildcards like x* are no longer applied to IDNA names (which start with 'xn--') - fix crash of Utils::CERT_free - support TLSv11, TLSv12 as handshake protocols 1.955 2013/10/11 - support for forward secrecy using ECDH, if the Net::SSLeay/openssl version supports it. 1.954 2013/9/15 - accept older versions of ExtUtils::MakeMaker and add meta information like link to repository only for newer versions. 1.953 2013/7/22 - fixes to IO::Socket::SSL::Utils, thanks to rurban[AT]x-ray[DOT]at, RT#87052 1.952 2013/7/11 - fix t/acceptSSL-timeout.t on Win32, RT#86862 1.951 2013/7/3 - better document builtin defaults for key,cert,CA and how they are depreceated - use Net::SSLeay::CTX_set_default_verify_paths to use openssl's builtin defaults for CA unless CA path/file was given (or IO::Socket::SSL builtins used) 1.950 2013/7/3 - MAJOR BEHAVIOR CHANGE: ssl_verify_mode now defaults to verify_peer for client. Until now it used verify_none, but loudly complained since 1.79 about it. It will not complain any longer, but the connection might probably fail. Please don't simply disable ssl verification, but instead set SSL_ca_file etc so that verification succeeds! - MAJOR BEHAVIOR CHANGE: it will now complain if the builtin defaults of certs/my-ca.pem or ca/ for CA and certs/{server,client}-{key,cert}.pem for cert and key are used, e.g. no certificates are specified explicitly. In the future these insecure (relative path!) defaults will be removed and the CA replaced with the system defaults. v1.94 2013.06.01 - Makefile.PL reported wrong version of openssl, if Net::SSLeay was not installed instead of reporting missing dependency to Net::SSLeay. v1.93 2013.05.31 - need at least OpenSSL version 0.9.8 now, since last 0.9.7 was released 6 years ago. Remove code to work around older releases. - changed AUTHOR in Makefile.PL from array back to string, because the array feature is not available in MakeMaker shipped with 5.8.9 (RT#85739) v1.92 2013.05.30 - Intercept: use sha1-fingerprint of original cert for id into cache unless otherwise given - Fix pod error in IO::Socket::SSL::Utils RT#85733 v1.91 2013.05.30 - added IO::Socket::SSL::Utils for easier manipulation of certificates and keys - moved SSL interception into IO::Socket::SSL::Intercept and simplified it using IO::Socket::SSL::Utils - enhance meta information in Makefile.PL v1.90 2013.05.27 - RT#85290, support more digest, especially SHA-2. Thanks to ujvari[AT]microsec[DOT]hu - added support for easy SSL interception (man in the middle) based on ideas found in mojo-mitm proxy (which was written by Karel Miko) - make 1.46 the minimal required version for Net::SSLeay, because it introduced lots of useful functions. v1.89 2013.05.14 - if IO::Socket::IP is used it should be at least version 0.20, otherwise we get problems with HTTP::Daemon::SSL and maybe others (RT#81932) - Spelling corrections, thanks to dsteinbrunner v1.88 2013.05.02 - consider a value of '' the same as undef for SSL_ca_(path|file), SSL_key* and SSL_cert* - some apps like Net::LDAP use it that way. Thanks to alexander[AT]kuehn[AT]nagilum[DOT]de for reporting the problem. v1.87 2013.04.24 - RT#84829 - complain if given SSL_(key|cert|ca)_(file|path) do not exist or if they are not readable. Thanks to perl[AT]minty[DOT]org - fix use of SSL_key|SSL_file objects instead of files, broken with 1.83 v1.86 2013.04.17 - RT#84686 - don't complain about SSL_verify_mode is SSL_reuse_ctx, thanks to CLEACH v1.85 2013.04.14 - probe for available modules with local __DIE__ and __WARN__handlers. fixes RT#84574, thanks to FRAZER - fix warning, when IO::Socket::IP is installed and inet6 support gets explicitly requested. RT#84619, thanks to Prashant[DOT]Tekriwal[AT]netapp[DOT]com v1.84 2013.02.15 - disabled client side SNI for openssl version < 1.0.0 because of RT#83289 - added functions can_client_sni, can_server_sni, can_npn to check availability of SNI and NPN features. Added more documentation for SNI and NPN. v1.83_1 2013.02.14 - separated documentation of non-blocking I/O from error handling - changed and documented behavior of readline to return the read data on EAGAIN/EWOULDBLOCK in case of non-blocking socket. See https://github.com/noxxi/p5-io-socket-ssl/issues/1, thanks to mytram v1.83 2013.02.03 - Server Name Indication (SNI) support on the server side, inspired by patch provided by karel[DOT]miko[AT]gmail[DOT]com. https://rt.cpan.org/Ticket/Display.html?id=82761 - reworked part of the documentation, like providing better examples. v1.82 2013.01.28 - sub error sets $SSL_ERROR etc only if there really is an error, otherwise it will keep the latest error. This causes IO::Socket::SSL->new.. to report the correct problem, even if the problem is deeper in the code (like in connect) - correct spelling, rt#8270. Thanks to ETHER v1.81 2012.12.06 - deprecated set_ctx_defaults, new name ist set_defaults (but old name still available) - changed handling of default path for SSL_(ca|cert|key)* keys: either if one of these keys is user defined don't add defaults for the others, e.g. don't mix user settings and defaults - cleaner handling of module defaults vs. global settings vs. socket specific settings. Global and socket specific settings are both provided by the user, while module defaults not. - make IO::Socket::INET6 and IO::Socket::IP specific tests run both, even if both modules are installed by faking a failed load of the other module. v1.80 2012.11.30 - removed some warnings in test (missing SSL_verify_mode => 0) which caused tests to hang on Windows. https://rt.cpan.org/Ticket/Display.html?id=81493 v1.79 2012.11.25 - prepare transition to a more secure default for SSL_verify_mode. The use of the current default SSL_VERIFY_NONE will cause a big warning for clients, unless SSL_verify_mode was explicitly set inside the application to this insecure value. In the near future the default will be SSL_VERIFY_PEER, and thus causing verification failures in unchanged applications. v1.78 2012.11.25 - use getnameinfo instead of unpack_sockaddr_in6 to get PeerAddr and PeerPort from sockaddr in _update_peer, because this provides scope too. Thanks to bluhm[AT]genua[DOT]de. - work around systems which don't defined AF_INET6 https://rt.cpan.org/Ticket/Display.html?id=81216 Thanks to GAAS for reporting v1.77 2012.10.05 - update_peer for IPv6 also, applied fix to https://rt.cpan.org/Ticket/Display.html?id=79916 by tlhackque[AT]yahoo[DOT]com v1.76 2012.06.18 - no longer depend on Socket.pm 1.95 for inet_pton, but use Socket6.pm if no current Socket.pm is available. Thanks to paul[AT]city-fan[DOT]org for pointing out the problem and providing first patch v1.75 2012.06.15 - made it possible to explicitly disable TLSv11 and TLSv12 in SSL_version v1.74_2 2012.06.07 - fixed documentation errors, reported by MARSCHAP https://rt.cpan.org/Ticket/Display.html?id=77690 v1.74_1 2012.06.07 - add support to IO::Socket::IP which support inet6 and inet4 by integrating patch from PEVANS for https://rt.cpan.org/Ticket/Display.html?id=75218 v1.74 2012.05.13 - accept a version of SSLv2/3 as SSLv23, because older documentation could be interpreted like this v1.73 2012.05.11 - make test t/dhe.t hopefully work for more version of openssl Thanks to paul[AT]city-fan[DOT]org for providing bug reports and testing environment v1.72 2012.05.10 - set DEFAULT_CIPHER_LIST to ALL:!LOW instead of HIGH:!LOW Thanks to dcostas[AT]gmail[DOT]com for problem report v1.71 2012.05.09 - 1.70 done right. Also don't disable SSLv2 ciphers, SSLv2 support is better disabled by the default SSL_version of 'SSLv23:!SSLv2' v1.70 2012.05.08 - make it possible to disable protocols using SSL_version, make SSL_version default to 'SSLv23:!SSLv2' v1.69 2012.05.08 - re-added workaround in t/dhe.t v1.68 2012.05.07 - remove SSLv2 from default cipher list, which makes failed tests after last change work again, fix behavior for empty cipher list (use default) v1.67 2012.05.07 - https://rt.cpan.org/Ticket/Display.html?id=76929 thanks to d[DOT]thomas[AT]its[DOT]uq[DOT]edu[DOT]au for reporting - if no explicit cipher list is given it will now default to ALL:!LOW instead of the openssl default, which usually includes weak ciphers like DES. - new config key SSL_honor_cipher_order and documented how to use it to fight BEAST attack. v1.66 2012.04.16 - make it thread safer, thanks to bug report from vega[DOT]james[AT]gmail [DOT]com, https://rt.cpan.org/Ticket/Display.html?id=76538 v1.65 2012.04.16 - added NPN (Next Protocol Negotiation) support based on patch from kmx https://rt.cpan.org/Ticket/Display.html?id=76223 v1.64 2012.04.06 - clarify some behavior regarding hostname verification. Thanks to DOHERTY for reporting. v1.63 2012.04.06 - applied patch of DOUGDUDE to ignore die from within eval to make tests more stable on Win32, https://rt.cpan.org/Ticket/Display.html?id=76147 v1.62 2012.03.28 - small fix to last version v1.61 2012.03.27 - call CTX_set_session_id_context so that servers session caching works with client certificates too. https://rt.cpan.org/Ticket/Display.html?id=76053 v1.60 2012.03.20 - don't make blocking readline if socket was set nonblocking, but return as soon no more data are available https://rt.cpan.org/Ticket/Display.html?id=75910 - fix BUG section about threading so that it shows package as thread safe as long as Net::SSLeay >= 1.43 is used https://rt.cpan.org/Ticket/Display.html?id=75749 v1.59 2012.03.08 - if SSLv2 is not supported by Net::SSLeay set SSL_ERROR with useful message when attempting to use it. - modify constant declarations so that 5.6.1 should work again v1.58 2012.02.26 - fix t/dhe.t again to enable the workaround only for newer openssl versions, because this would cause failures on older versions v1.57 2012.02.26 - fix t/dhe.t for openssl 1.0.1 beta by forcing tlsv1, so that it does not complain about the too small rsa key which it should not use anyway. Thanks to paul[AT]city-fan[DOT]org for reporting. https://rt.cpan.org/Ticket/Display.html?id=75165 v1.56 2012.02.22 - add automatic or explicit (via SSL_hostname) SNI support, needed for multiple SSL hostnames with same IP. Currently only supported for the client. v1.55 2012.02.20 - work around IO::Sockets work around for systems returning EISCONN etc on connect retry for non-blocking sockets by clearing $! if SUPER::connect returned true. https://rt.cpan.org/Ticket/Display.html?id=75101 Thanks for Manoj Kumar for reporting. v1.54 2012.01.11 - return 0 instead of undef in SSL_verify_callback to fix uninitialized warnings. Thanks to d[DOT]thomas[AT]its[DOT]uq[DOT]edu[DOT]au for reporting the bug and MIKEM for the fix. https://rt.cpan.org/Ticket/Display.html?id=73629 v1.53 2011.12.11 - kill child in t/memleak_bad_handshake.t if test fails https://rt.cpan.org/Ticket/Display.html?id=73146 Thanks to CLEACH ofr reporting v1.52 2011.12.07 - fix syntax error in t/memleak_bad_handshake.t thanks to cazzaniga[DOT]sandro[AT]gmail[DOT]com for reporting v1.51 2011.12.06 - disable t/memleak_bad_handshake.t on AIX, because it might hang https://rt.cpan.org/Ticket/Display.html?id=72170 v1.50 2011.12.06 Thanks to HMBRAND for reporting and Rainer Tammer tammer[AT]tammer[DOT]net for providing access to AIX system v1.49 2011.10.28 - another regression for readline fix, this time it failed to return lines at eof which don't end with newline. Extended t/readline.t to catch this case and the fix for 1.48 Thanks to christoph[DOT]mallon[AT]gmx[DOT]de for reporting v1.48 2011.10.26 - bugfix for readline fix in 1.45. If the pending data where false (like '0') it failed to read rest of line. Thanks to Victor Popov for reporting https://rt.cpan.org/Ticket/Display.html?id=71953 v1.47 2011.10.21 - fix for 1.46 - check for mswin32 needs to be /i. Thanks to Alexandr Ciornii for reporting v1.46 2011.10.18 - disable test t/signal-readline.t on windows, because signals are not relevant for this platform and test does not work. https://rt.cpan.org/Ticket/Display.html?id=71699 v1.45 2011.10.12 - fix readline to continue when getting interrupt waiting for more data. Thanks to kgc[AT]corp[DOT]sonic[DOT]net for reporting problem v1.44 2011.05.27 - fix invalid call to inet_pton in verify_hostname_of_cert when identity should be verified as ipv6 address, because it contains colon. v1.43_1 2011.05.12 - try to make t/nonblock.t more stable, especially on Mac OS X v1.43 2011.05.11 - fix t/nonblock.t - stability improvements t/inet6.t v1.42 2011.05.10 - add SSL_create_ctx_callback to have a way to adjust context on creation. https://rt.cpan.org/Ticket/Display.html?id=67799 - describe problem of fake memory leak because of big session cache and how to fix it, see https://rt.cpan.org/Ticket/Display.html?id=68073 v1.41 2011.05.09 - fix issue in stop_SSL where it did not issue a shutdown of the SSL connection if it first received the shutdown from the other side. Thanks to fencingleo[AT]gmail[DOT]com for reporting - try to make t/nonblock.t more reliable, at least report the real cause of ssl connection errors v1.40 2011.05.02 - integrated patch from GAAS to get IDN support from URI. https://rt.cpan.org/Ticket/Display.html?id=67676 v1.39_1 2011.05.02 - fix in exampel/async_https_server. Thanks to DetlefPilzecker[AT]web[DOT]de for reporting v1.39 2011.03.03 - fixed documentation of http verification: wildcards in cn is allowed v1.38_1 2011.01.24 - close should undef _SSL_fileno, because the fileno is no longer valid (SSL connection and socket are closed) v1.38 2011.01.18 - fixed wildcards_in_cn setting for http (wrongly set in 1.34 to 1 instead of anywhere). Thanks to dagolden[AT]cpan[DOT]org for reporting https://rt.cpan.org/Ticket/Display.html?id=64864 v1.37 2010.12.09 - don't complain about invalid certificate locations if user explicitly set SSL_ca_path and SSL_ca_file to undef. Assume that user knows what he is doing and will work around the problems by itself. http://rt.cpan.org/Ticket/Display.html?id=63741 v1.36 2010.12.08 - update documentation for SSL_verify_callback based on https://rt.cpan.org/Ticket/Display.html?id=63743 https://rt.cpan.org/Ticket/Display.html?id=63740 v1.35 2010.12.06 - if verify_mode is not VERIFY_NONE and the ca_file/ca_path cannot be verified as valid it will no longer fall back to VERIFY_NONE but throw an error. Thanks to Salvatore Bonaccorso and Daniel Kahn Gillmor for pointing out the problem, see also http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=606058 v1.34 2010.11.01 - scheme http for certificate verification changed to wildcards_in_cn=1, because according to rfc2818 this is valid and also seen in the wild - if upgrading socket from inet to ssl fails due to handshake problems the socket gets downgraded, but is still open. See https://rt.cpan.org/Ticket/Display.html?id=61466 - deprecate kill_socket, just use close() v1.33 2010.03.17 - attempt to make t/memleak_bad_handshake.t more stable, it fails for unknown reason on various systems - fix hostname checking: an IP should only be checked against subjectAltName GEN_IPADD, never against GEN_DNS or CN. Thanks to rusch[AT]genua[DOT]de for bug report v1.32 2010.02.22 - Makefile.PL: die if Scalar::Util has no dualvar support instead of only complaining. Thanks to w[DOT]phillip[DOT]moore[AT]gmail[DOT]com for reporting. v1.31 2009.09.25 - add and export constants for SSL_VERIFY_* - set SSL_use_cert if cert is given and not SSL_server - support alternative CRL file with SSL_crl_file thanks to patch of w[DOT]phillip[DOT]moore[AT]gmail[DOT]com v1.30_3 2009.09.03 - make t/memleak_bad_handshake.t more stable (increase listen queue, ignore errors on connect, don't run on windows..) v1.30_2 2009.09.01 - t/memleak_bad_handshake.t don't write errors with ps to stderr, -o vsize argument is not supported on all platforms, just skip test then v1.30_1 2009.08.31 - make sure that idn_to_ascii gets no \0 bytes from identity, because it simply cuts the string their (using C semantics). Not really a security problem because IDN like identity is provided by user in hostname, not by certificate. v1.30 2009.08.19 - fix test t/memleak_bad_handshake.t v1.29 2009.08.19 - fixed thanks for version 1.28 v1.28 2009.08.19 - fix memleak when SSL handshake failed. Thanks richardhundtu[AT]gmail[DOT]com v1.27 2009.07.24 - changed possible local/utf-8 depended \w in some regex against more explicit [a-zA-Z0-9_]. Fixed one regex, where it assumed, that service names can't have '-' inside - fixed bug https://rt.cpan.org/Ticket/Display.html?id=48131 where eli[AT]dvns[DOT]com reported warnings when perl -w was used. While there made it more aware of errors in Net::ssl_write_all (return undef not 0 in generic_write) v1.26 2009.07.03 - SECURITY BUGFIX! fix Bug in verify_hostname_of_cert where it matched only the prefix for the hostname when no wildcard was given, e.g. www.example.org matched against a certificate with name www.exam in it Thanks to MLEHMANN for reporting v1.25 2009.07.02 - t/nonblock.t: increase number of bytes written to fix bug with OS X 10.5 https://rt.cpan.org/Ticket/Display.html?id=47240 v1.24 2009.04.01 - add verify hostname scheme ftp, same as http - renew test certificates again (root CA expired, now valid for 10 years) v1.23 2009.02.23 - if neither SSL_ca_file nor SSL_ca_path are known (e.g. not given and the default values have no existing file|path) disable checking of certificates, but carp about the problem - new test certificates, the old ones expired and caused tests to fail v1.22 2009.01.24 - Net::SSLeay stores verify callbacks inside hash and never clears them, so set verify callback to NULL in destroy of context v1.21 2009.01.22 - auto verification of name in certificate created circular reference between SSL and CTX object with the verify_callback, which caused the objects to be destroyed only at program end. Fix it be no longer access $self from inside the callback. Thanks to odenbach[AT]uni-paderborn[DOT]de for reporting v1.20 2009.01.15 - only changes on test suite to make it ready for win32 (tested with strawberry perl 5.8.8) v1.19 2008.12.31 - fix verifycn_name autodetection from PeerAddr/PeerHost v1.18 2008.11.17 - fixed typo in argument: wildcars_in_cn -> wildcards_in_cn http://rt.cpan.org/Ticket/Display.html?id=40997 thanks to ludwig[DOT]nussel[AT]suse[DOT]de for reporting v1.17 2008.10.13 - no code changes, publish v.16_3 as v.17 because it looks better than v.16 - document win32 behavior regarding non-blocking and timeouts v1.16_3 2008.09.25 - fix t/nonblock.t with workaround for problems with IO::Socket::INET on some systems (Mac,5.6.2) where it cannot do nonblocking connect and leaves socket blocked. - make some tests less verbose by fixing diag in t/testlib.t (send output to STDOUT not STDERR and prefix with '#') v1.16_2 2008.09.24 - work around Bug in IO::Socket::INET6 on BSD systems http://rt.cpan.org/Ticket/Display.html?id=39550 by setting Domain based on PeerAddr Thanks to srezic for report and support - remove tests of recv/send from t/core.t. Might badly interact with SSL handshake and cause crashes as seen on OS X 10.4 v1.16_1 2008.09.19 - better support for IPv6: - IPv6 is enabled by default if IO::Socket::INET6 is available - t/inet6.t for basic tests v1.16 2008.09.19 - change code for SSL_check_crl to use X509_STORE_set_flags instead of X509_STORE_CTX_set_flags based on bug report from <tjtoocool[AT]phreaker[DOT]net > - change opened() to report -1 if the IO::Handle is open, but the SSL connection failed, needed with HTTP::Daemon::SSL which will send an error message over the unencrypted socket v1.15 - change internal behavior when SSL handshake failed (like when verify callback returned an error) in the hope to fix spurious errors in t/auto_verify_hostname.t v1.14 - added support for verification of hostname from certificate including subjectAltNames, support for IDN etc based on patch and input from christopher[AT]odenbachs[DOT]de and achim[AT]grolmsnet[DOT]de. It is also possible to get more information from peer_certificate based on this patch. See documentation for peer_certificate and verify_hostname - automatic verification of hostnames with SSL_verifycn_scheme and SSL_verifycn_name - global setting of default context options like SSL_verifycn_scheme, SSL_verify_mode with set_ctx_defaults - fix import of inet4,inet6 which got broken within 1.13_X. Thanks to <at[AT]altlinux[DOT]ru> for bugreport and patch - clarified and enhanced debugging support based on bugreport http://rt.cpan.org/Ticket/Display.html?id=32960 - put information into README regarding the supported and recommended version of Net::SSLeay v1.13 - removed CLONE_SKIP which was added in 1.03 because this breaks windows forking. Handled threads/windows forking better by making sure that CTX from Net::SSLeay gets not freed multiple times from different threads after cloning/forking - removed setting LocalPort to 0 in tests, instead leave it undef if a random port should be allocated. This should fix build problems with 5.6.1. Thanks to <andrew[DOT]benham[AT]thus[DOT]net> v1.12 - treat timeouts of 0 for accept_SSL and connect_SSL like no timeout, like IO::Socket does. v1.11 - fixed errors in accept_SSL which would work when called from start_SSL but not from accept v1.10 - start_SSL, accept_SSL and connect_SSL have argument for Timeout so that the SSL handshake will not block forever. Only used if the socket is blocking. If not set the Timeout value from the underlying IO::Socket is used v1.09 - new method stop_SSL as opposite of start_SSL based on an idea of Bron Gondwana <brong[AT]fastmail[DOT]fm> To support this method the SSL_shutdown handling had to be fixed, e.g. in close a proper unidirectional shutdown should be done while in stop_SSL a bidirectional shutdown - try to make it clearer that thread support is buggy v1.08 - make sure that Scalar::Util has support for dualvar (Makefile.PL,SSL.pm) because the perl-only version has has no dualvar v1.07 - fix t/nonblock.t on systems which have by default a larger socket buffer. Set SO_SNDBUF explicitly with setsockopt to force smaller writes on the socket v1.06 - instead of setting undef args to '' in configure_SSL drop them. This makes Net::SMTP::SSL working again because it does not give LocalPort of '' to IO::Socket::INET any more v1.05 - make session cache working even if the IO::Socket::SSL object was not created with IO::Socket::SSL->new but with IO::Socket::SSL->start_SSL on an established socket v1.04 - added way to create SSL object with predefined session cache, thus making it possible to share the cache between objects even if the rest of the context is not shared key SSL_session_cache Note that the arguments of IO::Socket::SSL::SessionCache::new changed (but you should never have used this class directly because it's internal to IO::Socket::SSL) v1.03 - add CLONE_SKIP as proposed by Jarrod Johnson jbjohnso at us dot ibm dot com v1.02 - added some info to BUGS and to BUGS section of pod - added TELL and BINMODE to IO::Socket::SSL::SSL_HANDLE, even if they do nothing useful. - all tests allocate now the ports dynamically, so there should be no longer a conflict with open ports on the system where the tests run v1.01 - work around Bug in Net::HTTPS where it defines sub blocking as {}, e.g. force scalar context when calling sub blocking (in IO::Socket::SSL::write) see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=383106 v1.0 - fix deprecated and practically undocumented function get_peer_certificate so that LWP Net::HTTPS works again - set arg 'Blocking' while calling SUPER::configure only if it was set by the caller to work around Problem in LWP Net::HTTPS v0.999 - If SSL_cipher_list is not given it uses the openssl default instead of setting it to 'ALL:!LOW:!EXP' like before. The old value included ADH and this might be a bad idea, see BUGS why. v0.998 - declare socket as opened before calling fatal_ssl_error because the SSL_error_trap set up from HTTP::Daemon needs this - accept_SSL sets errors on $socket (the accepted socket) not $self (the listening socket if called from accept) so it can be queried from SSL_error_trap - note in BUGS section that IO::Socket::SSL is not thread-safe v0.997 - fix readline (e.g. getline,getlines,<>) so that it behaves regarding $/ like written in the $/ documentation. v0.996 - removed links and comments to unofficial release of Net::SSLeay, because there is a newer version already v0.995 - add support for Diffie Hellman Key Exchange. See parameter SSL_dh_file and SSL_dh. v0.994 - hide DEBUG statements and remove test to load Debug.pm because packages like SpamAssassin cannot cope with it (at least the OpenBSD port) v0.993 - added SSL_cert and SSL_key parameter which do not take a file name like SSL_cert_file and SSL_key_file but an internal X509* resp. EVP_PKEY* value. Useful for dynamically created certificates and keys. - added test for sysread/syswrite behavior (which was changed in v0.991) v0.992 - _set_rw_error does $!||=EAGAIN only if error is one of SSL_WANT_READ|SSL_WANT_WRITE (patch from Mike Smith <mike at mailchannels dot com>) - Fix Makefile.PL to allow detection of failures in PREREQ_PM (http://rt.cpan.org/Public/Bug/Display.html?id=20563, patch by alexchorny at gmail dot com) v0.991 - sysread and syswrite ar no longer the same as read and write, but can return already if only parts of the data are read or written (which is the usual semantic for sysread and syswrite) This should fix problems with HTTP::Daemon::SSL v0.99 - just upgrade Version number because I've screwed up upload of v0.98 to cpan v0.98 - Maintainer changed to <Steffen_Ullrich at genua dot de> - Better support for nonblocking sockets: . exports $SSL_ERROR which contains the latest error from the openssl library. Exports constants SSL_WANT_READ and SSL_WANT_WRITE es special errors which will be set if openssl wants to write or read during nonblocking connects, accepts, reads or writes. . accept,accept_SSL,connect and connect_SSL don't block anymore if the socket is nonblocking. Instead $! will be set from the underlying IO::Socket::INET connect or accept if it failed there (usually EAGAIN or EINPROGRESS) or if the underlying openssl needs to read or write $! will be set to EAGAIN and $SSL_ERROR will be set to SSL_WANT_READ or SSL_WANT_WRITE . syswrite returns undef and sets $!,$SSL_ERROR if it fails to write instead of returning 0. - Bugfixes (http://rt.cpan.org/Public/Bug/Display.html?id=Bugid) . Bug 18439: fileno 0 should be valid . Bug 15001: sysread interprets buffer "0" as "" - peer_certificate returns X509 struct string if no field for extraction was specified - get_peer_certificate returns the certificate instead of the IO::Socket::SSL object v0.97 - Writes now correctly return errors. (Problem noted by Dominique Quatravaux <dom at idealx.com>). - CA paths now work without passing an empty SSL_ca_file argument. (Problem found by Phil Pennock, <phil.pennock at globnix.org>). - IO::Socket::SSL now automatically passes Proto => tcp (if not already specified) to IO::Socket::INET to work around /etc/services files with udp entries listed first. (Fix suggested by Phil Pennock). - $socket->accept() now returns the peer address in array context for better conformance with IO::Socket::INET. However, if you were doing "map { $_->accept } (@sockets)", or similar tricks, you will need to use "scalar" to get the old behavior back. (Problem noted by Nils Sowen, <n.sowen at kon.de>). - IO::Socket::SSL should now properly block on reads larger than the buffer size of Net::SSLeay. (Problem found by Eric Jergensen, <eric at dvns.com>). - IO::Socket::SSL should now send CA Certs (if necessary) along with certificates. (Problem found by <roy at momentous.ca>). - Timeouts should now work, but be aware that if multiple reads/writes are necessary to complete a connection, then each one may have a separate timeout. (Request from Dominique Quatravaux <dom at idealx.com>). - In certain cases, start_SSL() would misplace a socket's fileno, causing problems with starting SSL. This should now be fixed. (Problem found by <russ at zerotech.net>). - IO::Socket::SSL now requires a minimum of Net::SSLeay 1.21. --- Old Versions -------------------------------------------------- v0.96 2004.4.30 - Makefile's error messages now correct if output is redirected (patch from Ilya Zakharevich <ilya at math.berkeley.edu>). - Non-blocking connects/accepts now work (Problem found by Uri Guttman <uri at stemsystems.com>). - new_from_fd() now works. - getline() and <> in scalar context now return undef instead of '' if the read failed. (Problem found by Christian Gilmore <cag at us.ibm.com>). - Broken pipe signals are now ignored during socket close to prevent a SSL shutdown message from killing the parent program. (Problem found by Christian Gilmore). - Tests should proceed much more quickly, and a semi-race was fixed, meaning that on slow machines the tests should be more reliable. - Check for Scalar::Util and Weakref now uses default $SIG{__DIE__} instead of a potentially user-altered one (suggestion from Olaf Schneider <Olaf.Schneider at iwr.fzk.de>). This only applies to Perl 5.6.0 & above. - Session caching support (patch from Marko Asplund <marko.asplund at kronodoc.fi>). - set_default_context() added to alter the behavior of modules that use IO::Socket::SSL from the main program. - get_ssl_object() renamed to _get_ssl_object() to reflect the fact that it's only supposed to be used internally (not that you should have cared, of course). - Added patch for Net::SSLeay to take advantage of client-side session caching. v0.95 2003.8.25 - Changed PeerAddr in example/ssl_client.pl back to localhost. - Update of examples to automatically switch to the proper directory if they cannot find the necessary SSL certificates. - Minor documentation update with more INET6 info. - Corrected some error messages for IO::Socket::INET6. - Better opened() behavior when sockets close unexpectedly. - Added note about random number generators for Solaris users (Problem found by Christian Gilmore <cag at us.ibm.com>). - Added support for WeakRef and Scalar::Util to allow IO::Socket::SSL objects to auto-destroy themselves when they go out of scope. - Added croak()ing for unimplemented send() and recv() methods so they are not accidentally used to transmit unencrypted data. The Perl builtin functions cannot be reliably trapped and are still dangerous, a fact that the POD now reflects (Problem noted by Michal Ludvig <michal at logix.cx>). v0.94 2003.6.26 - Changed accept() to use inherited accept() instead of IO::Socket::accept, so that IPv6 inheritance is possible. - Added options to import() so that a user could specify IPv6 or IPv4 mode of operation. - Documentation fixes, esp. e-mail address. v0.93 2003.6.24 - Fixed error-checking slip in connect_SSL() (Problem found by Uri Guttman <uri at stemsystems.com>). - All functions now return the empty list () on errors. - Added note about the above change to appease Graham Barr <gbarr at pobox.com>. - Fixed Net::SSLeay giving warnings when arguments are undef; in all cases, undef arguments may be set to '' without any change in behavior except for removing the warnings. (Problem found by Dominique Quatravaux <dom at idealx.com>) - If accept() or connect() fails in SSL negotiation, the user now has the option to print something to the failed socket before it is closed. (error_trap option in new()) - Added support for CRLs (SSL_check_crl option in new()) for versions of OpenSSL >= 0.9.7b (Original patch from Brian Lindauer <jbl at sysd.com>) - Finally added decent support for certificate callbacks. (SSL_verify_callback option in new(), suggestion from Dariush Pietrzak <eyck at ghost.anime.pl>). - accept()/connect()/socket_to_SSL() now fail immediately if the socket in question does not have a fileno. - Added the kill_socket() method to guarantee that a socket dies. - Fixed extra warning when printing errors in debug mode. - Deprecated socket_to_SSL() in favor of the class method start_SSL() (Class method suggestion from Graham Barr <gbarr at pobox.com>). - Added the class method start_SSL() to allow for cases when the desired class of the socket is not IO::Socket::SSL (Request from Dariush Pietrzak <eyck at ghost.anime.pl>) - Changed socket_to_SSL to rebless socket to original class if SSL negotiation failed (Request from Graham Barr <gbarr at pobox.com>) - Removed the daemon.pl example, as it did not work with the standard distribution of HTTP::Daemon (use HTTP::Daemon::SSL instead). v0.92 2002.10.22 - Changed the fileno() function to support returning the fileno of server sockets. (Problem found by Roland Giersig <RGiersig at cpan.org>). - Fixed SSL_version incorrectly defaulting to SSLv2 (patch from Roland Alder <roland.alder at celeris.ch>). v0.91 2002.08.31 - Added support for SSL_peek and SSL_pending (peek() and pending()). Updated documentation, tests, etc. to reflect this. v0.901 2002.08.19 - Fixed the warning that happens when sockets are not explicitly closed() before the program terminates. v0.90 2002.08.13 - This version is a complete rewrite of IO::Socket::SSL. It now has about half the lines of code, twice the amount of documentation, and a slightly more polished interface. - IO::Socket::SSL now works properly with mod_perl and taint mode. - Major documentation update. - Update of the BUGS file to reflect changes made in the rewrite. - Update of the test suite for Perl v5.8.0 (or, more precisely, for Scalar::Util). - Update of the test suite for Perl v5.00503 (or, more precisely, for the lack of several nice features added in v5.6.0) (Marko Asplund <aspa at kronodoc.fi>). - New test suite that does not need the Internet to function. - Update of all the files in example/ to use more current features of IO::Socket::SSL. - Removal of SSL_SSL and X509_Certificate classes. - There have been a few name changes (like socketToSSL -> socket_to_SSL) for better consistency. - The functionality of get_peer_certificate() and friends is deprecated. - The functionality of want_write() and want_read() is deprecated. - The functionality of context_init() is deprecated for normal use. - Support for all SSL context options in the new() call. - SSL contexts are no longer global. The SSL_reuse_ctx option is provided for those who want to re-use a context. - The default verify mode is now VERIFY_NONE. - IO::Socket::SSL::DEBUG is now linked to Net::SSLeay::trace to provide different levels of debugging information. - There is a uniform interface for error reporting, so on error all functions will return undef and the error will be available by calling errstr(). - The dump_peer_certificate() and peer_certificate() functions have been added. - sysread() will now behave correctly if the offset argument is greater than the length of the read buffer. It also will truncate the read buffer properly, according to the Perl documentation for sysread(). - getline(), getlines(), and getc() have been added. - syswrite() now uses references to avoid copying large amounts of data. - readline() uses ssl_read_all in array context for improved speed. - close() now uses SSL_shutdown() to properly close an SSL connection, unless you tell it not to. - If you have Net::SSLeay version 1.18 or greater, X509 certificates will be properly freed. - All other known bugs have been fixed. v0.81a (Not publicly released) - Added support for SSL_passwd_cb. - Added accept() server socket support to socketToSSL(). v0.81 2002.04.10 - calling context_init twice destroyed global context. fix from Jason Heiss <jheiss at ofb.net>. - file handle tying interface implementation moved to a separate class to prevent problems resulting from self-tying filehandles. Harmon S. Nine <hnine at netarx.com>. - docs/debugging.txt file added - require Net::SSLeay v1.08 - preliminary support for non-blocking read/write - socketToSSL() now respects context's SSL verify setting reported by Uri Guttman <uri at stemsystems.com>. v0.80 2001.08.19 - fixed startTLS support (socketToSSL) (Graham Barr <gbarr at pobox.com>) - make accept() set fileno attribute on newly created IO::Socket::SSL object (Martin Oldfield <m at mail.tc>). - certificate updates. - use SSL_CTX_use_PrivateKey_file in SSL_Context::new. v0.79 2001.06.04 - angle bracket readline operator support (David Darville <david at dark.x.dtu.dk>). - eliminate warnings in choosing SSL protocol version. - implement our own opened method and make length parameter optional in syswrite (Robert Bihlmeyer <robbe at orcus.priv.at>). v0.78 2001.04.24 - test script targets changed, certificate setup fixed - support for TLS in SSL_version. SSL_version parameter values changed from integer to string. NB: this is an incompatible change. all SSL_version parameter values have to be changed. valid values include: 'sslv2', 'sslv3', 'sslv23'. Stephen C. Koehler <koehler at securecomputing.com>. - enable selecting SSL version for connections. patch from Takanori Ugai <ugai at jp.fujitsu.com>. - allow setting SSL_ca_file to ''. this is needed for being able to use SSL_ca_path (Robert Bihlmeyer <robbe at orcus.priv.at>). - include the Apache CA bundle file in the distribution (my-ca.pem). - BUGS file added. v0.77 2001.01.15 - don't setup SSL CA verification unless cert verification is actually used for the connections. - default SSL protocol version selection in SSL.pm. v0.76 2000.11.17 - patch from Kwok Chern Yue <chernyue at post1.com> for making IO::Socket::SSL work with HTTP::Daemon. v0.75 2000.07.26 - IO::Socket::SSL should now work with perl v5.6.0 - demo/*.pl and t/*.t now turn module debugging on if DEBUG command line argument is given - default certificates changed v0.74 2000.07.05 - Changes file added - bugfix in IO::Socket::SSL::sysread() (zliu2 at acsu.buffalo.edu) - libwww-perl and IO::Socket::SSL UML models added in docs - URL changes in test scripts - preliminary support for startTLS in IO::Socket::SSL::socketToSSL() - miscellaneous patches for Net::SSLeay added in diffs